AI-First Test Automation · Open Source · Apache 2.0

Ship Faster With
AI-Powered Testing

SentinelFlux generates test suites from your knowledge base, heals broken locators automatically, and analyzes every failure with AI — covering API, Web, Security, Performance, and more from a single dashboard.

View on GitHub
$ pip install sentinelflux
localhost:8000 — SentinelFlux Dashboard
Test Runs
247
↑ 94.3% pass rate
AI Agents
11
Post-suite analysis
Test Functions
169
AI-generated
Doc Coverage
100%
All scripts documented
Recent Test Runs
orangehrm · web
28/28 passed
restfulbooker · api
13/13 passed
reportportal · vapt
3 findings
AI Pipeline
DocGendone
DocReviewdone
ScriptGendone
ScriptReviewrunning

Everything a QA team needs, in one framework

Eight testing modules, one CLI, one dashboard. No juggling between tools.

REST & GraphQL
Schema validation, data-driven tests, full CRUD + curl log on failure
Web UI Testing
Page Object Model with 3-tier self-healing locators via Playwright
Security (VAPT)
OWASP A01–A10, 31 checks, PDF report + compliance certificate
Performance
Load, stress, spike, soak — p50/p75/p95/p99 metrics with thresholds
Accessibility
WCAG 2.1 A/AA/AAA via axe-core, per-page violations by impact
API Contract
Validates real responses against your OpenAPI spec — status + body schema
Visual Regression
Pixel-diff screenshots against baseline, configurable threshold alerts
Bug Tracker
Full defect lifecycle — state machine, artifacts, comments, PDF incident report
AI Test Pipeline

From Knowledge Base to
test suite, automatically

Define your product's forms, fields, APIs, and business rules in YAML. SentinelFlux's pipeline reads your KB, generates structured test case docs, reviews them for quality, then writes executable pytest scripts — without hallucinating fields that don't exist.

  • DocGen → DocReview → ScriptGen → ScriptReview pipeline
  • AppExplorerAgent crawls your live app for verified DOM selectors before generating
  • TC ID normalizer backstop prevents duplicate or out-of-sequence test IDs
  • Works with Mistral (cloud) or Ollama/Qwen (local, air-gapped)
run_pipeline.sh
# 1. Drop your KB files
ai/knowledge_base/myapp/
  application.yaml
  ui_pages.yaml
  api_specs.yaml

# 2. Explore + Generate
sentinelflux generate \
  --product myapp \
  --domain web \
  --explore \
  --base-url http://localhost:3000

# 3. Run
sentinelflux run --product myapp
Locator Healing Strategy
1
Primary Locator
data-testid, aria-label — most stable selectors
Active
2
Alternative Locators
id, name, placeholder, CSS path
Fallback
3
AI Healer
LocatorHealerAgent re-discovers from live DOM
Auto-heal
Human approval required before healing is committed
Self-Healing Locators

Tests that survive UI changes

When a locator breaks, SentinelFlux tries ranked alternatives before failing. If all fail, the LocatorHealerAgent re-discovers the element from the live DOM and queues a patch for human approval — so you fix broken tests, not broken code.

  • AppExplorerAgent pre-generates locators/web/*.json with real DOM selectors
  • Selectors ranked by stability: data-testid > id > name > aria-label
  • Approvals queue in the dashboard — AI suggests, you decide
Post-Suite AI Analysis

11 AI agents analyzing
every test run

After every run, SentinelFlux automatically classifies failures, detects flaky tests, guards against regressions, finds coverage gaps, and delivers strategic product insights — so you know exactly what needs attention before shipping.

ResultAnalyzer
FlakyDetector
RegressionGuard
CoverageGap
LocatorHealer
QuarantineManager
DocGen + DocReview
ScriptGen + Review
AppExplorer
CTOAgent
AI Analysis — After Run
test_login_invalid_credentials Product Bug

Error message not shown after invalid login — assertion failure on element visibility. Likely introduced in v2.3.1.

test_employee_search Flaky — Script

Failed 3 of 5 recent runs. Search input locator may need updating. Healing suggestion queued for approval.

Coverage Gap Detected 43 API scenarios

KB contains 43 API scenarios not covered by any test. Generate docs and scripts?

Up and running in three steps

Zero to AI-powered test automation, fast.

1

Install & Initialise

Install the framework and scaffold your product directory structure in one command.

$ pip install sentinelflux
$ sentinelflux init
2

Build Your Knowledge Base

Describe your application — forms, fields, APIs, business rules — in YAML. This is the AI's source of truth.

# ai/knowledge_base/myapp/
application.yaml
ui_pages.yaml
api_specs.yaml
3

Generate, Run & Analyze

AI writes your docs and scripts. Run from CLI or the dashboard. Agents analyze every failure automatically.

$ sentinelflux generate
$ sentinelflux run
# open localhost:8000

Open source. Free forever.

Apache 2.0 licensed. Star it, fork it, contribute to it.

Star on GitHub
$ pip install sentinelflux